This is one of those things that's hard to understand without practical moderation experience. The presence of an insulting meme creates the idiots who spam it, and creates a larger category of people who deploy it to toxify what would otherwise be polite and respectful discussion. And low quality comments that get a couple laugh reacts, even if you can consistently remove them within the hour, are fully capable of propagating it.

Keyword bans are definitely a heavy-handed option, they do risk the Streisand effect, and in the worst case that can require the scorched-earth counterresponse described in the source article. But sometimes there's just no other way to kill the meme.

They could kill the meme by...not producing slop.